There are many ways of making the coronavirus epidemic complicated. It is true that the Chinese account of what happened may be deficient, and that the numbers of deaths are probably underestimated. It is true that as each country received cases from China at somewhat different times and in different numbers it then went through its own particular epidemic, so comparing countries is subject to a timing bias.
Nonetheless, countries have their uses: they can close borders and establish a set of procedural rules and, to varying extents, implement them. This provides comparisons: will European countries do better than Taiwan, South Korea and Singapore? Will Sweden do better than the rest of Europe in terms of human cost vs lockdown cost? Are deaths per million the best comparison?
Complications arise because death comes with several precipitating causes, and death certificates must determine which of those finally finished the person off. John Graunt was aware of all these problems in 1662 when he published Natural and Political Observations Made upon the Bills of Mortality, thus launching demography and, at a very broad level, epidemiology. He made an astounding contribution. Here is a comparison I did of lifespans in his time and at the Millennium (shown for women only because I was discussing breast cancer).
As you can see, the population studied by Graunt began to thin out progressively from an early age. A few lived to the current lifespan, but nothing like the numbers of those alive today. An unrestricted pandemic would push us down somewhat, possibly similar to smoking being taken up again in large numbers.
Currently, dying “with”, but not “of” prostate cancer is a case in point as to the vagaries of classification, and the same applies to coronavirus as a cause of death. It is a bit like classifying deaths from influence A and B separately. To get round the problem of attribution, tracking excess deaths against seasonal averages is a good way of estimating what the real contribution of the pandemic may be.
A further complication is that it is hard to calculate how many people in the general population have been exposed to the virus. To do that properly, one would need a large random sample, perhaps 16,000, a highly accurate test, and a criterion group of previously infected patients, say 600 of them. Every test has an error term, even those touted as being very accurate. Even when it is trained on a criterion group there will be less than perfect sensitivity and specificity. Some of the signals the test picks up are ambiguous because they are borderline on a useful dimension, others can be misleading because an anti-body can have been triggered by other infections.
Then, when the test is used on the general public, the base rate problem rears its head. Small error rates in a test become very important when the disorder being tested for is rare. Population screening for breast cancer is an example. False positives are common, and are tolerated by doctors, if not so willingly by patients. Commonly, the chance that a positive test means that cancer is actually present may be only 10%. 9 patients get alarmed and are investigated for every one case found.
Some people react to all these complications by throwing up their hands and bemoaning that little is known about the virus. I demur. These are complications, but not fatal to drawing some preliminary conclusions. For example, this virus is not a Black Swan. It is yet another White Swan. (Strange as it may seem, I agree with Nassim Nicholas Taleb on this point). Much of what is known about coronaviruses applies to this recently encountered one. The contagion/fatality space is reasonably well known.
As a corrective to all these complications, here is a simplified version of the pandemic we face, and some possible solutions. It is intended to be simplistic.
It boils down to 2 factors: the viral load a person receives, and their immunological status.
Immunological status is particularly poor in old, ill and fat men. Shielding them would be a kindness, and should not require younger and fitter persons to be kept in idle captivity. The young shall inherit the earth, and the old will hold on to the capital, for the time being at least, as they dole it out to the young. Most of the elderly will survive, so long as they are not exposed to high viral loads.
Lockdowns provide a breathing space, reduce peak hospital demand, and temporarily diminish risks of infection for everyone. Any treatment intervention that saves lives (lives that the survivors themselves want to live) is to be welcomed, but it is more important to reduce viral loads in the first place.
High viral loads come principally from globules sprayed out by infected persons, as they sneeze, cough or in any way breathe heavily; secondarily from aerosols emitted by the same processes, least of all from either of these droplets lodged on surfaces. Getting from donor-nostril to recipient-nostril is the royal road for the virus. Energetic exhalations are the problem, hence dancing and singing and heavy exercises all create sprayed teaspoons full of goo which get breathed in by other people. Crowded and poorly ventilated rooms where such energetic activities happen create an infectious miasma. Sitting quietly in a room is less dangerous, and walking in the open air least dangerous of all, unless a succession of panting runners keep thundering close by.
When judging any activity, look at the distance between persons in still air, and estimate the degree of energetic exhalations. Clearly, wearing masks in enclosed places cuts down the spray of snot from the sneezer and the amount breathed in by the sneezed-upon. Any fabric will provide some reduction, N95 respirators a much greater reduction. Go the whole way and get a close-fitting neoprene mask with two permanent filters bolted on, and an escape valve so that you can breathe out directly, and the result is comfortable and effective.
In sum: wear facemasks when close to others in any confined space; provide good ventilation in all public buildings; and take part in plenty of open-air activities.